If one takes the power v. fuel flow graph for a given engine e.g. the IO-540 and extrapolates one of the constant-RPM lines to the zero-HP intercept, one finds several GPH fuel flow at that point. This fuel flow presumably represents the total of the pumping and friction losses, at that RPM. It is pretty significant, and explains why flying at a lower RPM, say 2200 v. 2400, does produce a significant improvement in MPG.
The engine should be more efficient when operating at a wide open throttle - because it doesn't have to suck quite as hard - but I have completely failed to measure any difference. Maybe I wasn't doing it right.
But there is no waste of fuel. In fact any non-turbo flight above about FL080 is done with a wide open throttle the whole time, and you just set the mixture for LOP or whatever grabs you. The RPM lever does give you a bit of power control: more RPM sucks in more air -> more fuel -> more power. At altitudes near the operating ceiling (20,000ft for me) one is at wide open throttle, max RPM (2575) and to hell with fuel consumption (beggars can't be choosers) so one sets it to 100F ROP.